Nowadays, with the continuous evolution of algorithms, chips, image sensing and other technologies, traffic monitoring is also developing in the direction of intelligence and high definition. With the gradual coverage of high-definition electronic police and road monitoring system, more vehicle information can be learned, more intelligent analysis ability is given to the monitoring system, and traffic violations are decreasing year by year. Hikvision is committed to bringing more perfect solutions for its intelligent transportation. The global shutter CMOS (hereinafter referred to as “GMOs”) camera of Hikvision series is one of the magic weapons.

The common COMS is roller shutter exposure, the exposure time of each line is the same, but the start time is different, which will lead to deformation and distortion of moving objects. GMOs Haikang series capture camera adopts global exposure mode, all pixel exposure can start and end at the same time, so the simultaneous exposure ensures the instantaneous shooting of high-speed moving objects.

Hikvision GMOs series capture machine adopts black gold paint metal shell, stable and generous, with a sense of massiness, all kinds of interfaces are also available. It is understood that the new generation of Hikvision Shenpi series products are closer to the needs of users and are committed to solving the practical problems of traffic users. For this kind of camera evaluation, we mainly consider three technical aspects: image quality performance, fluency and low illumination.

Eliminate the smile phenomenon from the root at the chip level

In the communication with the manufacturer, we know that the machine has black technology to ensure the image quality, and the picture has no small phenomenon. The way of CCD collecting and reading signals in specific cases, there will be charge overflow, leading to the emergence of upper and lower light bars, which is often called the “small phenomenon”. Hikvision uses GMOs as its sensor, and its single imaging point has completed the conversion from optical signal to digital signal, and the phenomenon of small phenomenon is eliminated from the chip layer.

Performance and evaluation of Hikvision GMOs series capture camera

High frame rate and attention to detail extraction

For traffic law enforcement, the higher the frame rate of the camera, the smoother the action per second, the stronger the sense of action, and the easier the accuracy of law enforcement. We evaluate the machine can reach 50 frames (full resolution). In the actual use process, the dual shutter technology can achieve both video brightness and capture rate. In the aspect of capture, based on artificial intelligence deep learning algorithm, the capture accuracy is greatly improved, and the structured data related to vehicle control, such as vehicle photo, license plate, vehicle model, logo, body color, can be easily extracted.

Double shutter effect

Still bright in the dark

Considering that urban road monitoring is an all-weather application scenario, it is required that the camera can clearly monitor or capture the specific information of the vehicle under various lighting conditions, and automatically analyze the vehicle information. By evaluating the GMOs chip scheme, the low illumination can be improved by about 15% (compared with the CCD scheme).

Ensure safety and promote smooth flow

In the process of traffic management from rough management centered on law enforcement to fine management centered on order management, information system construction needs more accurate, richer data base and more intelligent perception means. Hikvision GMOs series of capture machines help to innovate the means of traffic order management, and are the magic weapon of “ensuring safety and promoting smooth flow”.

The capture function of the camera is powerful, with the capture ability based on artificial intelligence, which greatly improves the quality of high-definition matting, and provides basic data for subsequent modeling and comparison. At the same time, this product of Hikvision supports the classification detection of non-human (motor vehicles, non motor vehicles and pedestrians) and the extraction of structured information such as vehicle feature recognition, providing accurate and valuable data for traffic order management, so as to realize the ultimate concept of “ensuring safety and promoting smooth traffic” traffic order management.

Manufacturer’s statement

This machine is the upgrade of Hikvision’s new generation of Intelligent Traffic Capture machine, which improves the function of the product. On the basis of the traditional electric police function, it integrates the N-in-One “super electric police” (n “12) of behavior detection such as disobedience to pedestrians, pedestrians running red lights, turning left but not straight, turning left but not right, turning left and turning small corner, and is increasing, And has launched a solution that can automatically capture and compare the driver’s face in the car. It is more possible to expand the application in the future.

Journalist’s point of view

GMOs is the source of black technology in this evaluation. It’s hard to imagine that Hikvision has not only achieved the best in the field of traditional snapshot machine, but also made great achievements in deep learning algorithm and photosensitive chip application. Equipped with this scheme, not only the traffic law enforcement will not appear the interference of smile phenomenon, the low illumination effect of night video is better, but also the full frame rate (25 frames / s) capture and video stream can be configured at the same time, providing more possibilities for the subsequent expansion of applications. GMOs solution in the performance of the machine, as for an industrial product equipped with a pair of brighter eyes; With the application of deep learning algorithm, this industrial product has a brain with stronger processing performance. Hikvision’s improvement in the above two aspects this year has laid a solid foundation for the realization of more user needs in the future.

Editor in charge: GT

Leave a Reply

Your email address will not be published. Required fields are marked *